How to add Python packages to object-detector-python? #43
-
|
I tried adding the following lines to Dockerfile: FROM python:3 It added about 500M to the image (not sure if there is a way to add just the packages I need), but then when I tried to start it using docker-compose, the following error message was displayed: object-detector-python-object-detector-python-1 | standard_init_linux.go:219: exec user process caused: exec format error |
Beta Was this translation helpful? Give feedback.
Replies: 3 comments 6 replies
-
|
Hello Igor, By adding |
Beta Was this translation helpful? Give feedback.
-
|
Sorry, one more question I added the following line to the Dockerfile in object-detector-python project: but when I try to build it, I still receive an error: Step 13/16 : RUN pip install RUN pip install Flask The Dockerfile seems to be identical with the one in pose-estimator-with-flask. The only difference that I see is the first line: ARG ARCH=aarch64 (in my Dockerfile) and ARG ARCH=armv7hf (in the pose-estimator-with-flask). Can it cause the error? |
Beta Was this translation helpful? Give feedback.
-
|
I think this happens because I am running build on linux/amd64 while requested platform is linux/arm64/v8 Not sure how to fix it. The step to install qemu to allow build flask for a different architecture produces a number of errors like this: I also tried docker buildx, but it also fails because it does not pick up proxy settings |
Beta Was this translation helpful? Give feedback.
Hello Igor,
By adding
FROM python:3ad the bottom of the file, you are building a docker image based on that, ignoring all the rest of the build. This will not work. You are going to need the libraries imported from the computer vison SDK.Besides, the image that you are pulling, "python:3" is an x86 image, which is not the right architecture for axis cameras (that is either armv7hf or aarch64).
If you need to add a python library you can simply do it without importing a new imaga, just like it is done in the pose-estimation-example.